416 S. Afr. J. Bot. 1997, 63(6) 416-420

Development and anatomy of the attachment structure of wood rose-producing mistletoes

1 1 C.M. Dzerefos ·2 and E.T.F. Witkowski* 1 Department of Botany, University of the Witwatersrand, Private Bag 3, Wits 2050 Republic of South Africa 2Wits Rural Facility, Private Bag X420, Acornhoek, 1360 Republic of South Africa

Received 25 Apri//997; revised 20Septeml>er /99"'

The formation of woodroses in Scferocarya birrea (A. Rich.) Hochst, and Combretum coffinum Fresen. in reaction to the parasitic mistletoes Erianthemum dregei (Eckl. & Zeyh .) Tiegh. and Pedisty/is gafpinii (Schinz ex Sprague) was investigated. Eight wood roses of varying sizes, and by implication age, were sectioned with the freeze microtome and the sledge microtome. As the mistletoe entered th e host tissues it assumed a flask-shape due to mechanical constriction by layers of host cork. There was a clear differentiation between host cells containing tannin and mistletoe cells without tannin throughout the sections. In addition these were separated by a thin wavy black hne of gum. The mistletoe occupied a central position as well as producing finger-like projections into the host tissue. Older sections indicated an increase in tannin and sclereids, while the arrangement of axial xylem elements became increasingly disrupted. There appear to be two strategies to prevent a functional connection with the mistletoe: 1) the production of wound periderm at the host surface and 2) internal necrosis and disruption of internal host tissues. The study species showed signs of both but the mistletoes managed to by-pass these strategies and instead lead to the formation of elaborate and intricate wood roses.

Keywords: Plant-parasite interactions; hypertrophy; mistletoes; haustorium; wound response.

•To whom correspondence should be addressed.

Introduction between January and August 1995. The P. galpinii individuals con­ Parasitic mistletoes obtain water and mineral solutes through a sisted of a one year old specimen germinated in July 1994 and a specialised organ of absorption, the haustorium, which pene­ smaller specimen, which was estimated to be about six months in 2 trates into the host to form a functional connection w ith the age having cross sectional areas of 33 and 63 rnm respectively. The xylem (Kuijt 1969; Fineran & Hocking 1983). The invading six E. dregei individuals were chosen on the basis of increasing size haustorium causes host tissue breakdown, which provides i.e. cross sectional areas 79. 3 534, 8 443, 12 920, 18 583, and 80 2 resources for mistletoe growth and leaves a space for the mistle­ 346 mm toe to occupy (Williams 1963; Briggs 1985). Hypertrophy (host Due to the range of woodrose sizes. different sectioning tc!ch­ tissue proliferation) can be a major host reaction (Menzies 1954; niques and storage methods were used to minimise specimen trag­ Weber 1993) and may result in an ornate proliferation of host tis­ mentation. P. galpinii spec imens were stored in a formalin-acetic sue called a woodrose (Figure 1). Other host responses have been alcohol (I'AA) solution (Briggs 1985). This was found to decrease documented. For example, Larix laricina (Du Roi) K. Koch. has fragmentation relative to specimens that had been stored dry. Trans­ been reported to produce wound periderm in its cortex, thus pre­ verse sections of the petiole and through the wood rose were made on an M SE microtome with a Pelcool cooling device, having a low tem­ venting the haustorium of the Yiscaceae mistletoe Arceuthobiwn perature range of 0° to 50°C and a high temperature range of 50° to pusillum Peck from reaching the host xylem (Tainter & French 100°C. Sections were cut at approximately 95°C while embedded in 1971 ). Tissue Tek OCT compound. a medium for frozen tissue specimens. The haustoria( anatomy of two mistletoe species Erianthemum Sections were stained on the slide with Sat"ranin-0 solution in 95% dregei (Eckl. & Zeyh.) Tiegh., formerly known as Loranthus alcohol for 10 min and differentiated with 95% alcohol until excess dregei Eckl. & Zeyh .• and the monotypic species Pedistylis gal­ stain was removed, followed by a quick rinse with absolute alcohol. pinii (Schinz ex Sprague) formerly known as Loranthus galpinii Sections were then countcrstained with fast green in 70% alcohol for Schinz (Wiens & Tolken 1979) were studied. Both have a wood­ 5 seconds and excess stain was removed with absolute alcohol. After rose type connection with their hosts, which has value in the a quick rinse with xylol, sections were mounted in DPX and covered curio trade (Dzerefos 1996). Thoday ( 1960) observed the onset with a cover slip. Observations and colour photographs were con­ ofwoodrose formation by E. dregei (on Burkea africana Hook.). ducted with a WILD 400 stereo microscope. H e observed that the host cambium rapidly proliferated when in Woodroses produced by the association of E. dregei on S. bl!'rea contact with the penetrating haustorium. The haustorium grew were stored dry in open. cardboard hoxe~. Three days after collec­ along the cambium, the margin of which became fluted such that tion the wood roses were cut transversely. into I 0 mm thick discs. the ti~sues of the host and the mistletoe interlocked. T he present with a circu lar saw, a quarter way from the host branch. Discs were investigation was conducted to determine suitable woodrose cut further with a handsaw into I 0 x I 0 mm strips and sanded to a growth monitoring methods and to gain an understanding of smooth finish by hand. Mistletoe tissue was distinctly weaker than woodrose formation. the host tissue and often crumbled during cutting. Specimens were stored in polytop tubes filled with tap water for two days. To Materials and Methods increase softening of the wood. the strips were placed into test tubes A total of eight woodroscs or different sizes, and by implication age, and gently heated to boiling point in a water bath. This process were sectioned to observe P. galpinii (two on Combretum collinum waterlogged the strips and caused them to swell. However. the dehy­ Fresen.) and £. dreKei [six on Sclerocwya birrea (A. Rich.) Hochst] dration process following sectioning reverted this to normal. infections. Specimens were collected from Klascrie (Mpumalanga) Strips were sectioned with a Reichert Sledge-Microtome at S. Afr. J. Bot. 1997. 63(6) 417

Figure 3 Ventral view of the six month old mistletoe Pedistylis galpinii surrounded by a collar of bract-like growths (PT. petiole: PH, primary haustorium) and the host Combretwn col/inwn (B. bract like growths: CR. cork). x 91. sections. Between tivc and ten sections were cut off each strip and kept in 70% ethanol. Sections were simultaneously taken through staining and dehydration in consecutiw petri-dishes containing: safranin-0 soluti on (30 min), distilled water (30 min). 50% ethanol (I 0 min}. 70% ethanol (10 min), 95% ethanol (10 min), 100% ethanol (10 min) and xylol (I 0 min). Two or three sections were mounted for each strip. Sections were straightened onto a glass slide, immersed with a few drops of the mountant Entellan R (Merck} and covered with a coverclip. Clothes pegs were used to fl atten the coverslip and sections. thus removing air-bubbles. Sections were studied at 40x and IOOx magnilication with a Nikon CFWE stereo microscope and black and white photo­ Figure l \Voodroscs ure the ornate. flowerlike remains of the graphs were taken. host after dcuth and removal of the mistletoe at the point of attach­ ment. This woodmse was formed by Pedisly/is galpinii on Sc/ero­ Results cw~va hirrca. Transverse sections (A) and (B) used for anatomical The six-month old association between the host, C. collinum, and investigations are indicated. the mistletoe, P. galpinii, showed two cotyledons emerging from a swollen, bark-covered holdfast within a developing woodrose FORESTEK (CSIR) in Pretoria. The strip was held by a vice on the consisting of host tissue. In comparison, the one year old P. microtome which advnnced automatically by 15 microns with each galpinii had eight leaves, was more swoll en, and had two distinct successive swipe of the blade. Sections were supported and mois­ zones of old and new bark. Longitudinal sections through the tened with a paintbrush dipped in 70% ethanol during the sectioning holdfast and the woodrose showed a flask-shaped mistletoe, with process. The n:mnants of the strips were dried at room temperature a swollen portion above the host branch. The mistletoe appeared· since the surface features and shape allowed orientation of the cut to be mechanically constricted by layers of cork as it entered the

Figure 2 Longitudinal section through a one year old woodrose Figure 4 Transverse section (A} indicating the transition zone constricted by layers of bark as it enters the host (P, the parasite between the six month old mistletoe Pcdistylis galpinii and Com­ Pedislylis galpinii: H. host branch: A. abnormal host outgrowth: 0, bretum collinum (CR. cork: T. tannin-tilled host cells: PH. primary uld bark: N. new bark: T. tannin-fi ll ed parenchyma). x 53. haustorium).x 46. 418 S. Afr. J. Bot. 1997, 63(6)

Discussion This histological investigation suggests woodrose-producing hosts employ two strategies to prevent functional connections with mistletoes. The firs t preventative measure involved the pro­ duction of wound periderm (i.e. cork, cork cambium and phello­ derm) at the host surface to provide mechanical resistance to the haustorium. Secondly, necrosis of internal host tissues and the increase of tyloses, lign in, tann in and gum isolate the haustorium from living host cell s. These strategies have varying levels of success which are host dependent. In the case of S. bin·ea and C. collinum both preventative strategi es were observed but they were not adequate to resist infection. In contrast, the host Junipe­ ms occidenta/is Hook. restricted penetration by the mistletoe Arceuthobium campylopodum Engelm . through both these strate­ gies (Jaramillo 1980), while other hosts res isted infection with only one strategy (Menzies 1954; Yan 1993). Figure 5 Transverst.: section (A) through l'edistylis galpinii A host wound periderm response was first visible in the six ~hmving four central vascular bundl t.:s (BR, bract-like growths; CR, and twelve month old mistletoes as a co llar- like outgrowth com­ cork; PR. parenchyma; T, tannin-filled parenchyma; V. vascular posed mostly of cork around the neck of the haustorium (Figure bundle). x 97. 3). The co llar was attributed to the mechanical pushing of the primary haustorium while entering the peripheral host tissues. tannin-tilled host tissues (Figure 2). Abnormal host growth was apparent as a finger-like projection growing through the cork Such an outgrowth was also observed in Loranthus micranthus layers. Under the dissecting microscope a collar consisting of Hook., and consisted of protruding epidermal and cortical layers bract-like growths was observed around the mi stletoe petiole (Menzies 1954). Swell ing at the mistletoe-host interface was (Figure 3) and may have been derived from host tissue. The ven­ probably due to the abnormal proliferation of host cambium tral view of the younger specimen had a tiny, grey-white area (Menzies 1954 ). Hoffman et a/. ( 1986 ), revealed, through ana­ between the petioles, composed mostly of conducti ng tissues. tomical investigations, that the penetration of the haustorium of This was evident in all complete transverse sections and was Tristerix tetrandus (R. et Pav.) Barlow et Wiens was unable to identified previously by Menzies ( 1954) as the primary hausto­ establish on non-host species due to the formation of wound peri­ rium. Transverse sections through the woodrose had an irregular derm. In Eucalyptus oleosa (infected by rlmyema preissii (Miq .) circumference (Figure 4). In unstained sections, outer host cell s Tiegh.) and in Heterodendrum ole(fa/iwn Desf. (infec ted by containing tannin appeared yellow-brown. In contrast, central Lysiana exocarpi (Behr.) Tiegh.) wound periderm formation mistletoe tissues were grey-white and without tannin. This dif­ usually, but not always, prevented the haustorium from invading ferentiation of host and mistletoe tissue was evident throughout further than the outermost bark of the host (Yen 1993). Similarly, the secti ons. The cortex appeared to be a zone of transition where wound periderm formation in the cortex of the host L. larcillia host tissue was folding in, and mi stletoe tissue was fo lding out. iso lated parts of the endophytic system of the Viscaceae mistle­ Islands of cork tissue within the host tissue were observed. toe A. pus ilium (Tainter & French 197 1). At low magnification the transverse section of the mistletoe A second strategy to thwart the mistletoe invasion was the petiole base showed four vascular bundles with xy lem, a cap of abnormal production of tyloses, lignin, tannin and gum. In S. bir­ phloem and a wide cortical region with parenchyma, freq uently rea wood tyloses were normally present (Kromhout 1977) but filled with tannin. The cork cambium or phellogen, and the corky not as prolific as seen in the wood rose sections (S . Dyer pers. layers of the collar were also observed (Figure 5). In unstained com.). An abundance oftyloses in host ax ial xylem elements pre­ transverse sections nearest the junction, mistletoe and host tissue vents water and mineral conduction (Raven et at. 1986) (Figure were clearly differentiated. The tannin-filled host cells appeared 7) and cou ld reduce or prevent loss of resources to the mistletoe. ye ll ow-brown and the tannin-less mistletoe cells were Tyloses are typically fo rmed during water stress. in leaf grey-white. The mi stletoe occupied a central position as well as producing finger-like projections into the host tissue (Figure 6). A dark grey wavy line of gum demarcated the boundary between host and mi stletoe. Longitudinal sections also showed gum at the interface which was discontinuous due to sectioning. There was a proliferation of tightly packed cambium ceiJs on the host side where the haustorium contacted the host. The host cortex was largely composed of groups of sclereids and parenchyma cells, which frequently contained tannin. This was surrounded by host cork interspersed with finger-like ingrowths. Xy lem was present in quantity but in spiralling disarray. Woodroses produced by E. dregei had an increase of tannin and sclereids in older sections relative to younger ones. There was an indication of exarch arrangement of xylem. Phloem could not be identified. The fin­ ger-like projections between the mistletoe and the host tissue occurred within the woodrose. In progressively older sections of woodroses, axial xylem elements seemed to become increasingly haphazard (Figure 7) and were orientated in at least two planes Figure 6 Transverse section (B) through the six month old asso­ on one section. Woodrose size, appeared to be directly related to ciation between Pedistylis galpinii and Combretum collinum . A an increasing number of tyloses blocking the axial xylem dark grey gum (G) demarcates tannin-tilled host cells (T) and tan­ elements . nin-less mistletoe cell s (P). x 11 5. S. Arr. J. Bot. 1997. 63(6) 419

Figure 7 Longitudinal tangential sections through: a) a small, b) a medium and c) a large woodrose produced by Erianthemum dregei on Scleromryu bin·ea (R. ray parenchyma; X, axial xylem clements blocked with tyloses). x 300.

absc iss ion zones (Cutter 1975) and after mechanical injury forms (Dzerefos 1996). Area of contact between mi stletoe and (Peters 1974). host was increased by the spiralling arrangement of axial xylem Host tissue lign ification increased with woodrose age, indicat­ elements (Figure 7) and invagination of haustoria! tissues (Figure ing a continual host reaction to infection. Menzies ( 1954) 6). A large contact area benefited mistletoes by providing a describes an abnormal 'gall ' produced at the mistletoe-host junc­ greater area of host tissue to access resources as well as a strong tion which was composed of lignified parenchyma cells. Abnor­ support. Woodrose connections were estimated to hold mistle­ mal wood of Pinus taeda L. also had increased lignifi cation toes weighing up to 100 kg (Weber 1993). Woodrose tissues (Herman 1988). were further strengthened by sclereids. £. dregei (on B. a.fricana Tannin production in the cortex also appeared to increase with and Lumnilzera racemosa Willd.) also had numerous sclereids woodrose age. Tannins in the bark of woody species prevented within the woodrose (Schonland 1907; Briggs 1985). germination and establishment of epiphytic orch ids (Frei & Dod­ Total woodrose growth occurs in both horizontal and vertical son 1972). The significance of tannin as a herbivore deterrent has planes and monitoring of growth should take both planes into been well documented (Crawley I 983; Du To it et al. 199 I; account. The convoluted margin of wood roses was indicative of Chesselet et a/. 1992) and is comparable to an anti-parasite the uneven growth activities along the margins of the hausto­ response as both resu lt in a loss of nutrients from the host. rium. Host cells were induced to divide both horizontally and A black gum was observed at the mistletoe-host interface and vertically, possibly allowing the haustorium to penetrate the probably consists of mistletoe-produced, digestive enzymes and softer tissues of the disorganised wood (Williams 1963). The dissolved host tissue (J. Kuijt pers. com.; Weber 1993). Menzies host cambium close to the haustorium of L. micranthus was also (1954) also observed a dark substance at the mistletoe-host inter­ noted to produce cell s orientated in all directions (Menzies face but attributed it to wounding. 1954). F. dregei and/'. galpinii manipu lated host resistance strategies by switching growth direction when host areas in direct contact Conclusion with the growing point of the haustorium become necrotic. In the At the cellular level host wound response to mistletoe infection mistletoe L. micranthus hausto ria! growth occurred in one direc­ was obvious due to the production of periderm, tyloses, lignin, tion then stopped while another portion in another direction tannin and gum . These substances increased with woodrose age resumed growth (Menzies 1954 ). This enabled contact between but apparently did not effect mistletoe persistence in the two host the actively growing haustorium and the new host tissues to be species investigated. Host xylem was in spirall ing disarray prob­ maintained and could account for the variety of woodrose growth ab ly making haustoria! penetration easier and increasing the sur- 420 S. Afr. J. Bot. 1997, 63(6)

face area in contact with the mistletoe axial xylem elements. P. COSTA, V. 1986. li-isterix tetrandus (Lnranthaccac) and its galpinii and E. dregei appear to be able to change growth direc­ host-plants in the Chi lean matomd: patterns and mechanisms. OLoranthaceae) in the Transvaal Lowveld. M.Sc. WEBER, H.C. 1993. Parasitismus von hiUtenpflanzen. pp. 57-511. dissertation, University of the Witwatersrand, Johannesburg. Dannstadt. Gennany. FINERAN, B.A. & HOCKING, P.J. 1983. Features of parasitism, mor­ WIENS, D. & TOLKEN. H. R. 1979. Loranthaceae. In: Flora of South­ phology and haustoria! anatomy in Loranthaceous root parasites. In: em Africa, cd. O.A. Leistner, Vol. 10. Botanical Research Institute. The Biology of Mistletoes. Chapter 12. pp. 205-227. Academic Press, Department of Agricultural Services, Pretoria, pp. 1-35. Australia. WILLIAMS, C.N. 1963. Development of Tapinantlms bangwensis FREI. J.K. & DODSON. C. H. 1972. The chemical effects of certain bark (Engler & Krause) Danser and contact with the host. Ann. Bot. 27: substances on the gem1ination and early growtl1 of epiphytic orchids. 64 1--646. Bull. Torrey Bot. Club. 99: 301-307. YAN, Z. 1993. Resistance to haustoria! development of two mistlctucs. HERMAN. R. 1988. Abnormal wood fonnation in Pinus taeda. PhD Amyema preissii (Miq. ) Tiegh. and l.ysiana e.rocarpi (Bchr.) Tiegh. Thesis. University of the Witwatersrand, Johannesburg. (Lomnthaceae), on host and nonlmst species. Int. J. Plant Sci. 154 : HOfFMAN. A.J .. FUENTES. E.R. CORTES, I., LIBERONA. F. & 386-394.